Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750921AbWIIKIW (ORCPT ); Sat, 9 Sep 2006 06:08:22 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751082AbWIIKIW (ORCPT ); Sat, 9 Sep 2006 06:08:22 -0400 Received: from dsl027-180-168.sfo1.dsl.speakeasy.net ([216.27.180.168]:52708 "EHLO sunset.davemloft.net") by vger.kernel.org with ESMTP id S1750921AbWIIKIV (ORCPT ); Sat, 9 Sep 2006 06:08:21 -0400 Date: Sat, 09 Sep 2006 03:08:54 -0700 (PDT) Message-Id: <20060909.030854.78720744.davem@davemloft.net> To: jeff@garzik.org Cc: paulus@samba.org, torvalds@osdl.org, linux-kernel@vger.kernel.org, benh@kernel.crashing.org, akpm@osdl.org, segher@kernel.crashing.org Subject: Re: Opinion on ordering of writel vs. stores to RAM From: David Miller In-Reply-To: <45028F87.7040603@garzik.org> References: <17666.11971.416250.857749@cargo.ozlabs.ibm.com> <20060909.023405.71099525.davem@davemloft.net> <45028F87.7040603@garzik.org> X-Mailer: Mew version 4.2 on Emacs 21.4 / Mule 5.0 (SAKAKI) Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 673 Lines: 15 From: Jeff Garzik Date: Sat, 09 Sep 2006 05:55:19 -0400 > As (I think) BenH mentioned in another email, the normal way Linux > handles these interfaces is for the primary API (readX, writeX) to be > strongly ordered, strongly coherent, etc. And then there is a relaxed > version without barriers and syncs, for the smart guys who know what > they're doing Indeed, I think that is the way to handle this. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/